CN115013258B - Intelligent soft measurement method for wind speed in front of wind turbine generator - Google Patents
Intelligent soft measurement method for wind speed in front of wind turbine generator Download PDFInfo
- Publication number
- CN115013258B CN115013258B CN202210796431.8A CN202210796431A CN115013258B CN 115013258 B CN115013258 B CN 115013258B CN 202210796431 A CN202210796431 A CN 202210796431A CN 115013258 B CN115013258 B CN 115013258B
- Authority
- CN
- China
- Prior art keywords
- wind speed
- wind
- turbine generator
- wind turbine
- model
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 238000000691 measurement method Methods 0.000 title claims abstract description 20
- 238000005259 measurement Methods 0.000 claims abstract description 23
- 230000003595 spectral effect Effects 0.000 claims abstract description 23
- 238000004422 calculation algorithm Methods 0.000 claims abstract description 22
- 230000009471 action Effects 0.000 claims abstract description 19
- 238000000354 decomposition reaction Methods 0.000 claims abstract description 19
- 238000004458 analytical method Methods 0.000 claims abstract description 14
- 238000013507 mapping Methods 0.000 claims abstract description 14
- 239000013598 vector Substances 0.000 claims abstract description 13
- 238000010801 machine learning Methods 0.000 claims abstract description 12
- 230000006870 function Effects 0.000 claims description 24
- 238000000034 method Methods 0.000 claims description 14
- 238000005096 rolling process Methods 0.000 claims description 6
- 238000005070 sampling Methods 0.000 claims description 6
- 230000007246 mechanism Effects 0.000 claims description 5
- 238000005311 autocorrelation function Methods 0.000 claims description 4
- 238000010219 correlation analysis Methods 0.000 claims description 4
- 230000008278 dynamic mechanism Effects 0.000 claims description 4
- 238000005457 optimization Methods 0.000 claims description 4
- 230000008569 process Effects 0.000 claims description 4
- 238000009826 distribution Methods 0.000 claims description 3
- 238000011156 evaluation Methods 0.000 claims description 3
- 238000004364 calculation method Methods 0.000 claims description 2
- 238000012545 processing Methods 0.000 claims description 2
- 238000007619 statistical method Methods 0.000 claims description 2
- 239000011295 pitch Substances 0.000 claims 1
- 238000012216 screening Methods 0.000 claims 1
- 230000008901 benefit Effects 0.000 abstract description 5
- 238000010276 construction Methods 0.000 abstract description 3
- 238000010248 power generation Methods 0.000 abstract description 3
- 238000013473 artificial intelligence Methods 0.000 abstract description 2
- 230000000694 effects Effects 0.000 description 7
- 238000010586 diagram Methods 0.000 description 4
- 108010076504 Protein Sorting Signals Proteins 0.000 description 2
- 230000008859 change Effects 0.000 description 2
- 238000004519 manufacturing process Methods 0.000 description 2
- 238000012935 Averaging Methods 0.000 description 1
- 238000013528 artificial neural network Methods 0.000 description 1
- 230000002457 bidirectional effect Effects 0.000 description 1
- 150000001875 compounds Chemical class 0.000 description 1
- 238000013135 deep learning Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000009699 differential effect Effects 0.000 description 1
- 230000007613 environmental effect Effects 0.000 description 1
- 230000006403 short-term memory Effects 0.000 description 1
- 238000004088 simulation Methods 0.000 description 1
- 238000001228 spectrum Methods 0.000 description 1
- 238000012706 support-vector machine Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01P—MEASURING LINEAR OR ANGULAR SPEED, ACCELERATION, DECELERATION, OR SHOCK; INDICATING PRESENCE, ABSENCE, OR DIRECTION, OF MOVEMENT
- G01P5/00—Measuring speed of fluids, e.g. of air stream; Measuring speed of bodies relative to fluids, e.g. of ship, of aircraft
- G01P5/26—Measuring speed of fluids, e.g. of air stream; Measuring speed of bodies relative to fluids, e.g. of ship, of aircraft by measuring the direct influence of the streaming fluid on the properties of a detecting optical wave
-
- F—MECHANICAL ENGINEERING; LIGHTING; HEATING; WEAPONS; BLASTING
- F03—MACHINES OR ENGINES FOR LIQUIDS; WIND, SPRING, OR WEIGHT MOTORS; PRODUCING MECHANICAL POWER OR A REACTIVE PROPULSIVE THRUST, NOT OTHERWISE PROVIDED FOR
- F03D—WIND MOTORS
- F03D17/00—Monitoring or testing of wind motors, e.g. diagnostics
-
- F—MECHANICAL ENGINEERING; LIGHTING; HEATING; WEAPONS; BLASTING
- F03—MACHINES OR ENGINES FOR LIQUIDS; WIND, SPRING, OR WEIGHT MOTORS; PRODUCING MECHANICAL POWER OR A REACTIVE PROPULSIVE THRUST, NOT OTHERWISE PROVIDED FOR
- F03D—WIND MOTORS
- F03D7/00—Controlling wind motors
-
- F—MECHANICAL ENGINEERING; LIGHTING; HEATING; WEAPONS; BLASTING
- F03—MACHINES OR ENGINES FOR LIQUIDS; WIND, SPRING, OR WEIGHT MOTORS; PRODUCING MECHANICAL POWER OR A REACTIVE PROPULSIVE THRUST, NOT OTHERWISE PROVIDED FOR
- F03D—WIND MOTORS
- F03D9/00—Adaptations of wind motors for special use; Combinations of wind motors with apparatus driven thereby; Wind motors specially adapted for installation in particular locations
- F03D9/20—Wind motors characterised by the driven apparatus
- F03D9/25—Wind motors characterised by the driven apparatus the apparatus being an electrical generator
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01S—RADIO DIRECTION-FINDING; RADIO NAVIGATION; DETERMINING DISTANCE OR VELOCITY BY USE OF RADIO WAVES; LOCATING OR PRESENCE-DETECTING BY USE OF THE REFLECTION OR RERADIATION OF RADIO WAVES; ANALOGOUS ARRANGEMENTS USING OTHER WAVES
- G01S17/00—Systems using the reflection or reradiation of electromagnetic waves other than radio waves, e.g. lidar systems
- G01S17/02—Systems using the reflection of electromagnetic waves other than radio waves
- G01S17/50—Systems of measurement based on relative movement of target
- G01S17/58—Velocity or trajectory determination systems; Sense-of-movement determination systems
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01S—RADIO DIRECTION-FINDING; RADIO NAVIGATION; DETERMINING DISTANCE OR VELOCITY BY USE OF RADIO WAVES; LOCATING OR PRESENCE-DETECTING BY USE OF THE REFLECTION OR RERADIATION OF RADIO WAVES; ANALOGOUS ARRANGEMENTS USING OTHER WAVES
- G01S17/00—Systems using the reflection or reradiation of electromagnetic waves other than radio waves, e.g. lidar systems
- G01S17/88—Lidar systems specially adapted for specific applications
- G01S17/95—Lidar systems specially adapted for specific applications for meteorological use
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N20/00—Machine learning
- G06N20/10—Machine learning using kernel methods, e.g. support vector machines [SVM]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/049—Temporal neural networks, e.g. delay elements, oscillating neurons or pulsed inputs
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02E—REDUCTION OF GREENHOUSE GAS [GHG] EMISSIONS, RELATED TO ENERGY GENERATION, TRANSMISSION OR DISTRIBUTION
- Y02E10/00—Energy generation through renewable energy sources
- Y02E10/70—Wind energy
- Y02E10/72—Wind turbines with rotation axis in wind direction
Abstract
The invention discloses an intelligent soft measurement method for the front wind speed of a wind turbine generator. Firstly, laser radar data is used as a data sample, AIC or BIC criteria are used for judging delay orders between input and output, dynamic differential regression vectors are defined according to the delay orders, a machine learning algorithm is used for dividing a hyperplane according to a clustering result, and an artificial intelligence dynamic regression algorithm is respectively adopted in a global action domain and a branch action domain for dynamic differential input and output nonlinear mapping modeling. Empirical mode decomposition and power spectral density analysis are carried out on the wind speed before the laser radar actual measurement machine and the wind speed before the soft measurement machine, so that the prediction performance of intelligent soft measurement of the wind speed before the wind turbine generator set is further verified. The invention has the advantages that: the laser radar system is reduced, the construction cost of the wind power plant is greatly reduced, the control performance of the unit can be optimized, the unit is ensured to operate safely and efficiently, and the power generation benefit of the wind power unit is improved.
Description
Technical Field
The invention relates to the technical field of wind power generation, in particular to an intelligent soft measurement method for the front wind speed of a wind turbine generator.
Background
With the continuous advancement of human society, the demand of electric power as an indispensable secondary energy in national production and life is becoming greater and greater. However, as the energy problem and the environmental problem become more severe, the gradual change of the traditional energy form to the new energy form has become a necessary trend in the development of the world. Among the renewable energy sources, wind energy is widely concerned due to its abundant resources and high efficiency.
When the wind turbine generator normally operates, due to the requirements of fan variable pitch control and output power prediction, the wind speed in front of the wind turbine generator is often required to be accurately acquired. Wind speed is generally collected by two methods at present: firstly, use the anemoscope of installing in fan cabin top to measure the wind speed, but because the anemoscope is located the wind turbine generator system cabin afterbody, and there is the certain distance between the wind wheel, so there is certain delay in the wind speed of surveying to because wake effect, the wind speed that the wind wheel rear position surveyed has the phenomenon of size and the frequent change of direction, leads to there being the deviation between cabin afterbody anemoscope surveys wind speed and the quick-witted preceding wind speed. And secondly, the laser radar wind measuring device is used for accurately acquiring the effective wind speed in front of the wind turbine, but because the laser radar is high in manufacturing cost, each fan of the wind power plant cannot be equipped.
In order to solve the problem, an intelligent soft measurement method for the wind speed in front of the wind turbine generator is provided. The method can successfully and softly measure the wind speed in front of the wind turbine generator system, and establishes an input and output nonlinear mapping model in regions based on a dynamic differential action region division idea. And performing soft measurement on the wind speed in front of the wind turbine generator according to the actually measured data of the laser radar and the established multiple nonlinear mapping models. The wind speed in front of the wind turbine can be accurately obtained, the control performance of the wind turbine can be optimized, the wind turbine can be safely and efficiently operated, and the power generation benefit of the wind turbine is improved.
Disclosure of Invention
In order to solve the various problems, the invention provides the intelligent soft measurement method for the wind speed in front of the wind turbine generator, which reduces the allocation of a laser radar system and reduces the construction cost of a wind power plant.
In order to solve the technical problems, the technical scheme provided by the invention is as follows: an intelligent soft measurement method for wind speed in front of a wind turbine generator comprises the following steps:
selecting a wind turbine generator, setting a sampling period, and acquiring actually measured operation data of the wind turbine generator by adopting an airborne or ground laser radar system;
according to the aerodynamic characteristic mechanism analysis of the wind turbine generator, selecting variables such as cabin wind speed and the like as model input variables, and selecting the wind speed in front of the wind turbine generator as model output variables;
analyzing the correlation between each input variable and each output variable based on a nonlinear feature selection algorithm, and finally determining an input feature variable;
determining a delay order existing between an input variable and an output variable by adopting a dynamic mechanism model and a Chichi Information Criterion, namely Akaike Information Criterion, AIC for short, or Bayesian Information Criterion, namely BIC for short, and defining a dynamic differential regression vector according to the delay order;
performing high-dimensional parameter machine learning clustering based on the dynamic differential regression vector, dividing a hyperplane based on a machine learning algorithm according to a clustering result, and obtaining a plurality of action domains according to the divided hyperplane;
uniformly selecting modeling samples in each action domain, establishing a global action domain dynamic differential input-output nonlinear mapping model and a sub-action domain dynamic differential input-output nonlinear mapping model by adopting an artificial intelligent dynamic regression algorithm, and performing soft measurement on the wind speed in front of the wind turbine by using rolling prediction;
and respectively carrying out empirical mode decomposition and power spectral density analysis on the radar measured wind speed and the soft measurement wind speed, comparing the global action domain model with the sub-action domain model, further verifying the prediction performance of the intelligent soft measurement method for the wind speed of the wind turbine generator set, and selecting the global action domain model or the sub-action domain model as a final soft measurement model according to the model performance.
Preferably, an airborne or ground laser radar system is used for ultra-short-term sampling of the wind turbine generator operation data.
Preferably, in order to verify the correlation between the input and output variables from a data plane, a maximum Information Coefficient method, namely, maximum Information Coefficient (MIC) for short, may be used to perform correlation analysis;
wherein, MIC has the formula
Preferably, the applied Chi-pool information criterion determines the order of delay that exists between input and output, which may be generally expressed as
Wherein n is a Representing the delay order of the output variable, n b And representing the delay order of the input variable, m representing the number of observations, and L representing the optimal value of the likelihood function obtained by the model.
Preferably, the adopted machine learning algorithm divides the hyperplane, and the optimization objective is as follows:
wherein, ω and b are normal vector and offset of the hyperplane respectively, C is a penalty factor, and the penalty strength to the misclassification sample is controlled by C, and the value range is (0,1). Zeta i Is a relaxation variable; y is i The value is 1 or-1 for the data classification label; m is the total amount of data used for classification.
And solving the optimization target to obtain the optimal division hyperplane coefficients of the adjacent data sets, wherein each hyperplane divides the whole scope into a plurality of sub scopes. The hyperplane is divided to clarify the boundary of each working condition, so that the working condition and the sample size of each working condition are conveniently identified, and modeling and model calling are facilitated.
Preferably, after empirical mode decomposition is performed on the radar measured wind speed and the model output wind speed, power is performed on the main empirical mode function, namely, intrinsic mode function, IMF for shortAnd (4) analyzing the spectrum density, and further verifying the effective wind speed prediction performance. The empirical mode decomposition is particularly suitable for analyzing and processing nonlinear non-stationary signals, and is a signal decomposition means, and the wind speed signals have the characteristics of nonlinear non-stationary and the like. The decomposition process is to find out all maximum value points of the original data sequence X (t) and fit the maximum value points by a cubic spline interpolation function to form an upper envelope curve of the original data. Similarly, all minimum value points are found out and are fitted through a cubic spline interpolation function to form a lower envelope line of the data, and the mean value of the upper envelope line and the lower envelope line is recorded as m l The average envelope m is subtracted from the original data sequence X (t) l Obtaining a new data sequence h, wherein the expression of the new data sequence h is
h=X(t)-m l (4)
And (3) subtracting new data after the envelope average from the original data, and if a negative local maximum value and a positive local minimum value exist, indicating that the new data is not an empirical mode function and the new data needs to be continuously screened.
The power spectral density, referred to as PSD, is the power carried by a unit frequency wave obtained by multiplying the power spectral density of a signal wave by a proper coefficient. The power spectral density function is an important statistical parameter of the frequency characteristic, and since the wind speed signal belongs to a random signal, the integral of which does not converge, the fourier transform of the wind speed signal itself does not exist, and therefore can be represented only by a statistical method. The autocorrelation function can completely represent the unique statistical average value of the signal, the power spectral density is the Fourier transform of the autocorrelation function, the power spectral density shows the power distribution condition of the signal at each frequency, and the maximum power output of the signal at a certain frequency can be judged through power spectral density analysis.
Compared with the prior art, the invention has the advantages that: wind measurement in the operation process of the wind turbine generator mainly depends on a cabin wind meter, a wind measuring tower and a laser radar system above the cabin. The data accuracy obtained by the laser radar system is highest, but the laser radar system is high in cost, so that the wind power plant cannot be equipped with the laser radar system for each fan, and only a few fans can be equipped with the laser radar. Therefore, the soft measurement of the wind turbine generator front wind speed is realized through the existing laser radar data and based on the wind turbine generator front wind speed intelligent soft measurement method, and the front wind speeds of other fans can be obtained through popularization of the laser radar data on a small number of fans. Therefore, the allocation of a laser radar system is reduced, and the construction cost of the wind power plant is greatly reduced.
The soft measurement of the wind speed in front of the wind turbine can improve the control performance of the wind generating set. Because the wind turbine is in a three-dimensional time-varying wind field environment, the wind speeds are distributed on the whole wind turbine rotation plane differently, and the wind speed measured by the anemometer is greatly different from the effective wind speed suffered by the whole wind turbine rotation plane, the wind speed before the wind turbine cannot be directly measured. However, the problem can be solved by an intelligent soft measurement method based on the wind turbine generator set front wind speed, because the front wind speed of the wind turbine generator can be indirectly obtained from other measurable data, and the accuracy is high.
The power characteristic of the wind turbine is an important index for evaluating the performance of the wind turbine, and the power characteristic directly influences the generated energy of the wind turbine. After the wind turbine generator is in service, power characteristic evaluation needs to be carried out in time, and the evaluation key lies in accurately acquiring the wind speed in front of the wind turbine generator. Therefore, the method has important guiding significance for the power characteristics of the wind turbine generator.
Drawings
FIG. 1 is a flow chart of an intelligent soft measurement method for wind speed in front of a wind turbine generator according to the present invention;
FIG. 2 is a K-means-hierarchical clustering effect diagram of the present invention;
FIG. 3 is a super-flat display effect diagram of the present invention;
FIG. 4 is a diagram of the structure of the BilSTM of the present invention;
FIG. 5 is a graph of input-output nonlinear mapping model accuracy of the present invention;
FIG. 6 is a graph of the rolling prediction effect of the present invention;
FIG. 7 is a (left) model output machine front wind speed EMD decomposition (right) actual measurement machine front wind speed EMD decomposition of the present invention;
FIG. 8 is a power spectral density plot of a pre-engine wind speed subsequence of the model of the present invention;
FIG. 9 is a power spectral density plot of a pre-line-in-time wind speed subsequence of the present invention.
Detailed Description
To make the objects, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings, and it is apparent that the described embodiments are some, but not all embodiments of the present invention. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
The technical means adopted by the invention to achieve the preset purpose are further described below by combining the drawings and the preferred embodiments of the invention. The method selects one wind generating set in a certain wind power plant in Yunnan, and uses a laser radar system to obtain actually-measured historical operating data for simulation, wherein the sampling period is 1 second/point.
Selecting a wind turbine generator, setting a sampling period, and acquiring actually measured operating data of the wind turbine generator by adopting an airborne or ground laser radar system;
secondly, selecting the wind speed of an engine room of the wind turbine generator, the rotating speed of a generator, active power and a pitch angle as model input variables and the wind speed in front of the wind turbine generator as model output variables according to the aerodynamic characteristic mechanism analysis of the wind turbine generator;
the wind turbine generator system aerodynamic characteristic mechanism analysis is characterized in that according to the aerodynamic principle, the power actually captured by a wind wheel after wind passes through the wind wheel is as follows:
wherein rho is air density, R is wind wheel radius, v q The wind speed before the aircraft, C p The wind energy utilization coefficient can be expressed as a nonlinear function between the tip speed ratio λ and the pitch angle β:
Cp=f(λ,β) (2)
wherein the tip speed ratio λ is defined asω r The rotational speed of the wind wheel rotor, the rotational speed of the wind wheel rotor and the rotational speed omega of the generator g The relationship between is ω g =αω r Where α is the gearbox ratio.
then, according to the law of conservation of energy, neglecting the mechanical side loss to obtain,
where m is the mass of gas flowing through, v c The nacelle wind speed.
So the wind speed v before the wind generator set q With generator speed omega g Cabin wind speed v c Power P and pitch angle beta, so the nacelle wind speed, generator speed, active power and pitch angle are selected as model input variables, and the pre-machine wind speed is selected as a model output variable.
And step three, analyzing the correlation between each input variable and each output variable by using a nonlinear correlation analysis algorithm, and finally determining the input characteristic variables.
In this embodiment, a Maximum Information Coefficient (MIC) method is used for correlation analysis, and MIC computation is divided into three steps:
1. giving i and j, meshing a scatter diagram formed by XY by i columns and j rows, and solving the maximum mutual information value;
2. normalizing the maximum mutual information value;
3. and selecting the maximum value of mutual information at different scales as the MIC value.
The correlation coefficient between each input variable and the output variable is shown in table 1.
Table 1 table of correlation coefficients between input variables and output variables
Step four, based on a dynamic mechanism model, judging a delay order existing between an input variable and an output variable by using an Akaike Information Criterion (AIC) or a Bayesian Information Criterion (BIC), and defining a dynamic differential regression vector according to the delay order;
in this embodiment, the AIC criterion is used to determine the order of delay between input and output as n a =3,n b =2, and the dynamic differential regression vector is defined according to the delay order as:
μ(t)=[g(t-1),g(t-2),P(t-1),P(t-2),v(t-1),v(t-2),p(t-1),p(t-2),
y(t-1),y(t-2),y(t-3)] (5)
wherein g (-) is the generator speed, P (-) is the active power, v (-) is the cabin wind speed, P (-) is the pitch angle, and y (-) is the pre-machine wind speed.
Classifying the data by adopting a high-dimensional parameter machine learning clustering algorithm based on the dynamic differential regression vector, dividing a hyperplane based on the machine learning algorithm according to a clustering result, and obtaining a plurality of action domains according to the divided hyperplane;
in this embodiment, the clustering algorithm adopts a K-means-hierarchical clustering algorithm, the machine learning algorithm adopts a soft-interval support vector machine algorithm, the clustering effect is shown in fig. 2, and the display effect of dividing the hyperplane is shown in fig. 3.
Step six, uniformly selecting modeling samples in each action domain, and respectively establishing an input and output nonlinear mapping model by adopting an artificial intelligence dynamic regression algorithm; based on the nonlinear mapping model of each scope of action, rolling prediction is used for realizing soft measurement of the wind speed in front of the wind turbine;
in this embodiment, a deep learning BilSTM neural network algorithm is used to build an input-output nonlinear mapping model, and a Bidirectional Long short term Memory (Bi-LSTM) algorithm is formed by combining a forward LSTM and a backward LSTM, because it combines information of an input sequence in both forward and backward directions. For output at time t, the forward LSTM layer has information of time t and previous times in the input sequence, and the backward LSTM layer has information of time t and later times in the input sequence. The vectors output by the two LSTM layers may be processed using addition, averaging, or concatenation. The Bi-LSTM structure is shown in FIG. 4. The modeling accuracy of the input-output nonlinear mapping model established based on one of the scopes is shown in fig. 5. The effect of soft measurements on pre-aircraft wind speed using rolling prediction is shown in figure 6.
And seventhly, performing empirical mode decomposition and power spectral density analysis on the radar measured wind speed and the soft measurement wind speed respectively, and further verifying the prediction performance of the wind speed intelligent soft measurement method before the wind turbine generator set.
The empirical mode decomposition comprises the following steps:
step 7.1, outputting a radar wind speed sequence l for the model p (t) and measured radar wind speed l (t), which will be described herein using measured radar wind speed l (t). Finding out all maximum points on the wind speed sequence, and forming an envelope curve l of the maximum points through a cubic spline interpolation function max (t), finding out all minimum value points on the wind speed sequence by the same method to form a minimum value point envelope curve l min (t); and (3) recording the mean value of the minimum value point envelope curve and the maximum value point envelope curve as a, and subtracting a from the wind speed sequence l (t) to obtain b (t), wherein the mean value is expressed as:
b(t)=l(t)-a (7)
step 7.2, consider b (t) as a new signal sequence, calculate its coefficient D k (ii) a Calculating the coefficient D k Is given by the formulaThe following:
where m is the number of data in the signal sequence, b k-1 (t)、b k (t) is the k-1 st calculation D k B (t) in coefficient.
Coefficient of judgment D k If it is between 0.1 and 0.2, if not b (t) is taken as l (t) and step 7.1 is repeated, after which the value of k is increased by 1 and step 7.2 is performed.
Step 7.3, if after k iterations satisfy D k Between 0.1 and 0.2, an empirical mode function is then obtained:
I n (t)=b k (t) (9)
wherein, I n (t) represents the nth empirical mode function.
Calculating the remainder of the nth empirical mode function:
ζ n (t)=l(t)-I n (t) (10)
step 7.4, judge I n (t) whether the function is a monotonic function or a constant, if yes, the decomposition is not completed, otherwise, let l (t) = ζ n (t), repeating the step 7.1 to the step 7.3, and continuously obtaining a new empirical mode function; all empirical mode functions and the remainder of the empirical mode functions obtained by the last decomposition are the radar wind speed subsequence.
The technical key point of the invention is how to construct an intelligent soft measurement method for the wind speed in front of the wind turbine generator. Firstly, laser radar data is used as a data sample, and an input variable and an output variable are determined through the analysis of a wind turbine generator aerodynamic characteristic mechanism. The method comprises the steps of judging the delay order between input and output by using an AIC (advanced information computer) or BIC (bit information computer) criterion based on a dynamic mechanism model, defining a dynamic differential regression vector according to the delay order, carrying out high-dimensional parametric machine learning clustering based on the dynamic differential regression vector, dividing a hyperplane by using a machine learning algorithm according to a clustering result, obtaining a plurality of action domains according to the hyperplane, carrying out input and output nonlinear mapping modeling by respectively adopting an artificial intelligent dynamic regression algorithm in each action domain, carrying out rolling prediction on the pre-machine wind speed based on the nonlinear mapping model of each action domain, and finally carrying out empirical mode decomposition and power spectral density analysis on the pre-machine wind speed of a laser radar actual measurement machine and the pre-machine wind speed of a model output machine, thereby further verifying the prediction performance of the pre-machine wind speed.
Here the model outputs a radar wind speed sequence l p And (t) the empirical mode decomposition process is the same as that of the actually measured radar wind speed l (t), and so on, and the detailed description is omitted. In this embodiment, the model outputs a radar wind speed sequence l p (t) are resolved with the measured radar wind speed l (t) to obtain 6 wind speed subsequences as shown in FIG. 7. And then carrying out power spectral density analysis on each decomposed wind speed subsequence, and searching a frequency range with the optimal wind speed soft measurement performance before the wind power station. The power spectral density of the model output radar wind speed subsequence is shown in fig. 8, and the power spectral density of the measured radar wind speed subsequence is shown in fig. 9.
The present invention and its embodiments have been described above, and the description is not intended to be limiting, and the drawings show only one embodiment of the present invention, and the actual structure is not limited thereto. In summary, those skilled in the art should appreciate that they can readily use the disclosed conception and specific embodiments as a basis for designing or modifying other structures for carrying out the same purposes of the present invention without departing from the spirit and scope of the invention as defined by the appended claims.
Claims (6)
1. An intelligent soft measurement method for wind speed in front of a wind turbine generator is characterized in that based on wind measurement of a laser radar in front of the wind turbine generator, dynamic differential short-term or ultra-short-term soft measurement of the wind speed in front of the wind turbine generator and performance evaluation of a time-frequency domain, the method comprises the following steps:
selecting a wind turbine generator, setting a sampling period, and acquiring actual measurement operation data of the wind turbine generator by adopting an airborne or ground laser radar system;
according to the aerodynamic characteristic mechanism analysis of the wind turbine generator, selecting variables such as cabin wind speed and the like as model input variables, and selecting the wind speed in front of the wind turbine generator as model output variables;
analyzing the correlation between each input variable and each output variable based on a nonlinear feature selection algorithm, and finally determining an input feature variable;
determining a delay order existing between an input variable and an output variable by adopting a dynamic mechanism model and a red pool Information Criterion, namely Akaike Information Criterion, AIC for short, or Bayesian Information Criterion, namely BIC for short, and defining a dynamic differential regression vector according to input and output variable pitches and the delay order thereof;
performing high-dimensional parameter machine learning clustering based on the dynamic differential regression vector, dividing a hyperplane based on a machine learning algorithm according to a clustering result, and obtaining a plurality of action domains according to the divided hyperplane;
uniformly selecting modeling samples in each action domain, establishing a global action domain dynamic differential input-output nonlinear mapping model and a sub-action domain dynamic differential input-output nonlinear mapping model by adopting an artificial intelligent dynamic regression algorithm, and performing soft measurement on the wind speed in front of the wind turbine by using rolling prediction;
and respectively carrying out empirical mode decomposition and power spectral density analysis on the radar measured wind speed and the soft measurement wind speed, comparing the global action domain model with the sub-action domain model, further verifying the prediction performance of the intelligent soft measurement method for the wind speed of the wind turbine generator set, and selecting the global action domain model or the sub-action domain model as a final soft measurement model according to the model performance.
2. The intelligent soft measurement method for the wind speed in front of the wind turbine generator set according to claim 1, characterized in that: and ultra-short-term sampling is carried out on the operation data of the wind turbine generator by using an airborne or ground laser radar system.
3. The intelligent soft measurement method for the wind speed in front of the wind turbine generator set according to claim 1, characterized in that: in order to verify the correlation between the input and output variables from a data layer, correlation analysis can be performed by adopting a maximum Information Coefficient method, namely Maximal Information Cooefficient, which is abbreviated as MIC;
4. the intelligent soft measurement method for the wind speed in front of the wind turbine generator set according to claim 1, characterized in that: the akage pool information criterion is used to determine the order of delay that exists between input and output, which can be generally expressed as:
wherein n is a Representing the order of delay, n, of the output variable b And representing the delay order of the input variable, m representing the number of observations, and L representing the optimal value of the likelihood function obtained by the model.
5. The intelligent soft measurement method for the wind speed in front of the wind turbine generator set according to claim 1, characterized in that: the adopted machine learning algorithm divides the hyperplane, and the optimization target is as follows:
wherein, ω and b are normal vector and offset of the hyperplane respectively, C is a penalty factor, and the penalty strength to the misclassification sample is controlled by C, and the value range is (0,1); zeta i Is a relaxation variable; y is i The value is 1 or-1 for the data classification label; m is the total amount of data used for classification;
solving the optimization target to obtain the optimal division hyperplane coefficients of the adjacent data sets, wherein each hyperplane divides the whole action domain into a plurality of sub-action domains; the hyperplane is divided to clarify the boundary of each working condition, so that the working condition and the sample size of each working condition are conveniently identified, and modeling and model calling are facilitated.
6. According to claimThe intelligent soft measurement method for the wind speed in front of the wind turbine generator set in the claim 1 is characterized in that: performing empirical mode decomposition on the actual measurement wind speed of the radar and the output wind speed of the model, and performing power spectral density analysis on a main empirical mode function, namely inrinsic mode function, IMF for short, so as to further verify the effective wind speed prediction performance; the empirical mode decomposition is particularly suitable for analyzing and processing nonlinear non-stationary signals, and is a signal decomposition means, and the wind speed signals have the characteristics of nonlinear non-stationary signals and the like; the decomposition process is to find out all maximum value points of the original data sequence X (t) and fit the maximum value points by a cubic spline interpolation function to form an upper envelope line of the original data; similarly, all minimum value points are found out and are fitted through a cubic spline interpolation function to form a lower envelope line of the data, and the mean value of the upper envelope line and the lower envelope line is recorded as m l The average envelope m is subtracted from the original data sequence X (t) l Obtaining a new data sequence h, wherein the expression of the new data sequence h is
h=X(t)-m l (4)
Subtracting new data after the envelope average from the original data, if a negative local maximum and a positive local minimum exist, indicating that the new data is not an empirical mode function, and needing to continue to carry out screening;
the power spectral density, namely Power Spectral Density (PSD), is the power carried by each unit frequency wave obtained by multiplying the power spectral density of the signal wave by a proper coefficient; the power spectral density function is an important statistical parameter of the frequency characteristic, and since the wind speed signal belongs to a random signal, the integral of the wind speed signal does not converge, so that the fourier transform of the wind speed signal does not exist per se, and therefore, the wind speed signal can be represented only by a statistical method; the autocorrelation function can completely represent the unique statistical average value of the signal, and the power spectral density is the Fourier transform of the autocorrelation function, which shows that the power distribution of the signal at each frequency is the distribution of the power of the signal, and the maximum power output of the signal at a certain frequency can be judged through power spectral density analysis.
Priority Applications (2)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202210796431.8A CN115013258B (en) | 2022-07-06 | 2022-07-06 | Intelligent soft measurement method for wind speed in front of wind turbine generator |
US18/126,027 US11789034B1 (en) | 2022-07-06 | 2023-03-24 | Intelligent soft measurement method for wind speed in front of wind turbine |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202210796431.8A CN115013258B (en) | 2022-07-06 | 2022-07-06 | Intelligent soft measurement method for wind speed in front of wind turbine generator |
Publications (2)
Publication Number | Publication Date |
---|---|
CN115013258A CN115013258A (en) | 2022-09-06 |
CN115013258B true CN115013258B (en) | 2022-12-09 |
Family
ID=83079614
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202210796431.8A Active CN115013258B (en) | 2022-07-06 | 2022-07-06 | Intelligent soft measurement method for wind speed in front of wind turbine generator |
Country Status (2)
Country | Link |
---|---|
US (1) | US11789034B1 (en) |
CN (1) | CN115013258B (en) |
Families Citing this family (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN116357522A (en) * | 2023-01-03 | 2023-06-30 | 蒙东协合开鲁风力发电有限公司 | Blower system integrated with airborne laser wind-finding radar and operation control method |
CN117744893A (en) * | 2024-02-19 | 2024-03-22 | 西安热工研究院有限公司 | Wind speed prediction method and system for energy storage auxiliary black start |
Family Cites Families (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
FR3013777B1 (en) * | 2013-11-25 | 2015-11-13 | IFP Energies Nouvelles | METHOD OF MONITORING AND MONITORING A WIND TURBINE USING WIND SPEED ESTIMATION USING A LIDAR SENSOR |
CN104653410A (en) * | 2014-09-10 | 2015-05-27 | 国家电网公司 | System and method for testing operation power characteristics of wind generation set based on laser radar wind meter |
US10803758B1 (en) * | 2016-12-21 | 2020-10-13 | Lockheed Martin Corporation | Wind prediction systems and methods |
JP2020153347A (en) * | 2019-03-22 | 2020-09-24 | 株式会社日立製作所 | Wind power generator |
CN111120205B (en) * | 2019-12-30 | 2021-01-26 | 明阳智慧能源集团股份公司 | Wind power plant regionalization control method based on laser radar |
EP3974646A1 (en) * | 2020-09-28 | 2022-03-30 | Ventus Engineering GmbH | Method and system for wind speed determination using vibration data |
-
2022
- 2022-07-06 CN CN202210796431.8A patent/CN115013258B/en active Active
-
2023
- 2023-03-24 US US18/126,027 patent/US11789034B1/en active Active
Also Published As
Publication number | Publication date |
---|---|
US11789034B1 (en) | 2023-10-17 |
CN115013258A (en) | 2022-09-06 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN115013258B (en) | Intelligent soft measurement method for wind speed in front of wind turbine generator | |
CN106873359B (en) | Wind power noise evaluation method based on cluster analysis and neural network | |
CN111091298B (en) | Wind power plant flow field coupling characteristic evaluation and intelligent grouping method and system | |
CN112594130A (en) | Machine learning model-based analysis for monitoring wind farm power performance | |
CN113313139A (en) | Wind power prediction uncertainty quantification method based on dynamic characteristics of unit | |
CN107045574A (en) | The low wind speed section effective wind speed method of estimation of wind power generating set based on SVR | |
CN113822418A (en) | Wind power plant power prediction method, system, device and storage medium | |
CN110991701A (en) | Wind power plant fan wind speed prediction method and system based on data fusion | |
CN114021483A (en) | Ultra-short-term wind power prediction method based on time domain characteristics and XGboost | |
Syu et al. | Ultra-short-term wind speed forecasting for wind power based on gated recurrent unit | |
CN115995810A (en) | Wind power prediction method and system considering weather fluctuation self-adaptive matching | |
CN110365053B (en) | Short-term wind power prediction method based on delay optimization strategy | |
CN114117887A (en) | Real-time assessment method, system and medium for online power generation performance of wind turbine generator | |
Asghar et al. | Online estimation of wind turbine tip speed ratio by adaptive neuro-fuzzy algorithm | |
CN111794921B (en) | Onshore wind turbine blade icing diagnosis method based on migration component analysis | |
CN113343562B (en) | Fan power prediction method and system based on hybrid modeling strategy | |
CN114320770A (en) | Effective wind speed measuring method and device for wind turbine generator | |
CN113610285A (en) | Power prediction method for distributed wind power | |
Cascianelli et al. | Experimental prediction intervals for monitoring wind turbines: An ensemble approach | |
CN113449920A (en) | Wind power prediction method, system and computer readable medium | |
CN113673786A (en) | Effective wind speed estimation system and method of wind driven generator based on RBF and LSTM | |
CN113309661B (en) | Method for improving generating capacity of wind generating set | |
CN116757101B (en) | Cabin wind speed correction method and system based on mechanism model and neural network | |
CN116822370B (en) | Ultra-short-term output prediction method for wind power cluster under data driving | |
Liu et al. | Wind-speed forecasting model based on DBN-Elman combined with improved PSO-HHT |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant | ||
GR01 | Patent grant |